基于词向量语义分类的微博实体链接方法-冯冲.pdf

《基于词向量语义分类的微博实体链接方法-冯冲.pdf》由会员分享，可在线阅读，更多相关《基于词向量语义分类的微博实体链接方法-冯冲.pdf（8页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、4261Vol.42, No.62016M6ACTA AUTOMATICA SINICA June, 2016M_lspL8ZE1F1Rt111;2K1pL8pMoV, 1H,34.3.YV9L8S&?L8ls+.4.T(10)9L8+F,iTKL8TKL8.TFlfi,5RS:NIL.3L3.1ylMV_alMV_aL8V#_M_(Wikipedia18,P2015M719S.YV?5Mo|id9,?V6U.V6L?Table 6 Scale of experiment data?lMVKey94293406lMVValue91948277lMVKey9213764lMVValue92354

2、687L894369348920142L|NLPCC20149L8_c177HpS400Hp(V|10000H|S)T_,V:9600Hp|100HST,kPNLPCC20142Z4k,c1152L8.3.2L!91)|rTzE1.YV1Er.s|NLPCC2014KZE(NLPCC)19_a/qL8ZE(EF*)20#Sc(CMEL)21ZE1.NLPCCSsL8MZEh, EF*q$FZE, CMELSL8:cTMZEL8h.LV,ZE(M).n5_sB,Kfi= 1:4, = 0:6,KkZE.|q(Precision)_azq(Recall)#F1TNS., in-KBsVUMoXlc

3、L8q, NILVUlcMoL8q.1LTV7V8U.V7 in-KBLTTable 7 Results of in-KBdqzqF1NLPCC 0.7927 0.8488 0.8198SCWE+EF 0.8137 0.8593 0.8358EF 0.7641 0.8142 0.7884CMEL 0.7951 0.8345 0.8143V8 NILLTTable 8 Results of NILdqzqF1NLPCC 0.9024 0.8653 0.8835SCWE+EF 0.9144 0.8763 0.8949EF 0.8871 0.8648 0.8758CMEL 0.8543 0.8694

4、 0.8461LT,Y8sVUZE.LV,ZEq_azqZ(AE,+YqA46.7ZENLPCC_aEF*CMELZE1uYZEFM_ls+,LT46VM_ls+r.,zQH*_a_a:_av-_aS_a%p_avX, NLPCCEF|v-,7ZE|-.2)1/M_ls+(SCWE)TY.SCWEL8+(EF)+,|LT1.TV9V10U.V9 in-KBLTTable 9 Results of in-KBqzqF10 0.7532 0.8016 0.77660.2 0.7621 0.8158 0.78800.4 0.7943 0.8375 0.81530.6 0.8137 0.8593 0.

5、83580.8 0.8032 0.8432 0.82271.0 0.7983 0.8488 0.8228V10 NILLTTable 10 Results of NILqzqF10 0.8432 0.8532 0.84820.2 0.8643 0.8713 0.86780.4 0.8917 0.8732 0.88240.6 0.9148 0.8762 0.89510.8 0.9032 0.8754 0.88911.0 0.9013 0.8743 0.8876VLTA, = 0H,ZE|L8T+,NHF1K.9, F199. = 0:6H, F1K. 0:6, F17S.VyL8ZErT46GM

6、_ls+.bVUF1W1,ym4.3)M_ls+k1.SCWEL8,|k4kW1.m5U,k= 10H,|KF1,515W, kMrqYv.k = 20H,SCWEF1v/.YVHpc_Md9,?CHp6:M_lspL8ZE921(7.91_M.sk = 20HF1/+|V,_.m4ZE/F1Fig.4 F1 scores of the combined measure withthe parameterm5 SCWEk/F1(Fig.5 F1 scores of SCWE with the k features4Zp_MMlbWL!,4M_lspL8lh,!9L8ZE,iNLPCC2014?

7、.LTVP4M_lsL8ZE,rTNLPCCX7KzT,qA46.T1,BM_mL8,=sL8.References1 Chinese Microblog Service. Sina Weibo User DevelopmentReport in 2014 Online, available: http:/ November 24, 2015(Sp_. 2014Mp?ZOnline, avail-able: http:/ Novem-ber 24, 2015)2 Guo Y H, Qin B, Liu T, Li S. Microblog entity linking byleveraging

8、 extra posts. In: Proceedings of the 2013 Confer-ence on Empirical Methods in Natural Language Process-ing. Seattle, USA: Association for Computational Linguis-tic, 2013. 8638683 Yang Jin-Feng, Yu Qiu-Bin, Guan Yi, Jiang Zhi-Peng. Anoverview of research on electronic medical record orientednamed ent

9、ity recognition and entity relation extraction.Acta Automatica Sinica, 2014, 40(8): 15371562(,l,1j,.0h_L8MYL81|8.1, 2014, 40(8): 15371562)4 Shen W, Wang J Y, Han J W. Entity linking with a knowl-edge base: issues, techniques, and solutions. IEEE Trans-actions on Knowledge and Data Engineering, 2015,

10、 27(2):4434605 Jiang L, Yu M, Zhou M, Liu X H, Zhao T J. Target-dependent twitter sentiment classiflcation. In: Proceedingsof the 49th Annual Meeting of the Association for Compu-tational Linguistics: Human Language Technologies. Port-land, Oregon, USA: 2011. 1511606 Shen W, Wang J Y, Luo P, Wang M.

11、 Linking named entitiesin tweets with knowledge base via user interest modeling.In: Proceedings of the 19th ACM SIGKDD InternationalConference on Knowledge Discovery and Data Mining. NewYork, USA: ACM, 2013. 68767 Liu X H, Li Y T, Wu H C, Zhou M, Wei F R, Lu Y. Entitylinking for tweets. In: Proceedi

12、ngs of the 51st Annual Meet-ing of the Association of Computational Linguistics. Sofla,Bulgaria: Association for Computational Linguistics, 2013.130413118 Odbal, Wang Zeng-Fu. Emotion analysis model using com-positional semantics. Acta Automatica Sinica, 2015, 41(12):21252137(r,9.BFlfs.1, 2015, 41(1

13、2): 21252137)9 NLPCC Online, available: http:/ sam.html. October 31, 201510 Hachey B, Radford W, Nothman J, Honnibal M, Curran JR. Evaluating entity linking with Wikipedia. Artiflcial In-telligence, 2013, 194: 13015011 Mikolov T, Chen K, Corrado G, Dean J. Ecient estimationof word representations in

14、 vector space. arXiv: 1301.3781,2013.12 Hartigan J A, Wong M A. Algorithm AS 136: a k-meansclustering algorithm. Journal of the Royal Statistical Soci-ety | Series C (Applied Statistics), 1979, 28(1): 10010813 Fernandez-Delgado M, Cernadas E, Barro S, Amorim D. Dowe need hundreds of classiflers to s

15、olve real world classifl-cation problems? Journal of Machine Learning Research,2014, 15: 3133318114 Mao Yi, Chen Wen-Lin, Guo Bao-Long, Chen Yi-Xin. Anovel logistic regression model based on density estimation.Acta Automatica Sinica, 2014, 40(1): 6272(j,!,RC,B.9 B.1, 2014, 40(1): 6272)92214215 Zhou

16、Xiao-Jian. Enhancing -support vector regressionwith gradient information. Acta Automatica Sinica, 2014,40(12): 29082915(kt.In0-_B.1, 2014,40(12): 29082915)16 King G, Zeng L C. Logistic regression in rare events data.Political Analysis, 2001, 9(2): 13716317 Guo Y H, Qin B, Li Y Q, Liu T, Lin S. Impro

17、ving can-didate generation for entity linking. In: Proceedings of the18th International Conference on Applications of NaturalLanguage to Information Systems. Salford, UK: Springer,2013. 22523618 Wikipedia Online, available: http:/zhwikilate-stzhwiki-latest-pages-articles.xml.bz2. Octo-ber 31, 201519

18、 Zhu Min, Jia Zhen, Zuo Ling, Wu An-Jun, Chen Fang-Zheng, Bai Yu. Research on entity linking of Chi-nese microblog. Acta Scientiarum Naturalium UniversitatisPekinensis, 2014, 50(1): 7378(,I,P*,-,Z,.pL8.v(1S), 2014, 50(1): 7378)20 Guo Yu-Hang. Research on Context-based Entity LinkingTechnique Ph.D. d

19、issertation, Harbin Institute of Tech-nology, China, 2014.(Rt./L8/pV,W:v,S, 2014.)21 Meng Z Y, Yu D, Xun E D. Chinese microblog entity link-ing system combining Wikipedia and search engine retrievalresults. In: Proceedings of the 3rd CCF Conference on Nat-ural Language Processing and Chinese Computi

20、ng. BerlinHeidelberg: Springer, 2014. 449456v9. 2005MSS/v9SpV.1Z_1),|,Jr.YT.E-mail: (FENG Chong Associate professorat the College of Computer Science andTechnology, Beijing Institute of Technology. He receivedhis Ph.D. degree from the Department of Computer Sci-ence, University of Science and Techno

21、logy of China in2005. His research interest covers natural language pro-cessing, information extraction, and machine translation.Corresponding author of this paper.)Fv9pV3.1Z_1),L8,sd.E-mail: (SHI Ge Ph.D. candidate at theCollege of Computer Science and Tech-nology, Beijing Institute of Technology.H

22、is research interest covers natural lan-guage processing, entity linking, and question answeringsystem.)Rtv9=.2014MW:v9S/pV.1Z_1),|,Jr.E-mail: (GUO Yu-Hang Lecturer at theCollege of Computer Science and Tech-nology, Beijing Institute of Technology.He received his Ph.D. degree from Harbin Institute o

23、fTechnology in 2014. His research interest covers naturallanguage processing, information extraction, and machinetranslation.)v9V3.1Z_1),Jr,sd.E-mail: (GONG Jing Master student at theCollege of Computer Science and Tech-nology, Beijing Institute of Technology.Her research interest covers natural lan

24、-guage processing, machine translation, and question an-swering system.)v9q.1989MSS9/9S/pV.1Z_1)Jr_,?)d.E-mail: (HUANG He-Yan Professor at theCollege of Computer Science and Tech-nology, Beijing Institute of Technology. She received herPh.D. degree from the Institute of Computing Technology,Chinese Academy of Sciences. Her research interest cov-ers natural language processing, machine translation, socialnetwork, information retrieval, and intelligent processingsystem.)